Expand document loader coverage (HTML, Markdown, DOCX, CSV, JSON, URL, S3, Image)

[galshubeli]

Problem

Only Text and PDF loaders ship today. Anyone with HTML, Markdown, DOCX, CSV/Excel, JSON, URL-based, S3-hosted, or image content has to write a custom LoaderStrategy before they can try the SDK.

Goal

Broaden built-in datatype support so common formats work out of the box.

Roadmap

Phase 1.1 (Document Loaders) and Phase 1.3 (Structured Data to Graph). Marked 🤝 — good for community contributions.

[drr00t]

@galshubeli I would like implement loaders for: DOCX, CSV/Excel and JSON. Use the Structural Chunk to create some kind of common format.

Is this make sense to you?

[galshubeli]

@galshubeli I would like implement loaders for: DOCX, CSV/Excel and JSON. Use the Structural Chunk to create some kind of common format.

Is this make sense to you?

yes, you want to take it?

[drr00t]

Yes, I would like to.